MCP HubMCP Hub
Retour aux compétences

resume

bitwize-music-studio
Mis à jour 2 days ago
7 vues
209
37
209
Voir sur GitHub
Autreai

À propos

Cette compétence reprend le travail sur un album musical existant en le localisant via une correspondance approximative du nom et en affichant son état détaillé avec les prochaines étapes. Elle est déclenchée lorsqu'un utilisateur mentionne un nom d'album ou souhaite poursuivre un travail précédent, utilisant les outils MCP pour localiser le projet et gérer les états du cache. La compétence fournit des chemins de récupération clairs si l'album n'est pas trouvé ou présente plusieurs correspondances.

Installation rapide

Claude Code

Recommandé
Principal
npx skills add bitwize-music-studio/claude-ai-music-skills -a claude-code
Commande PluginAlternatif
/plugin add https://github.com/bitwize-music-studio/claude-ai-music-skills
Git CloneAlternatif
git clone https://github.com/bitwize-music-studio/claude-ai-music-skills.git ~/.claude/skills/resume

Copiez et collez cette commande dans Claude Code pour installer cette compétence

Documentation

Resume Album Work

Purpose: Find an album and resume work where you left off.

Usage:

/bitwize-music:resume <album-name>
/bitwize-music:resume my-album
/bitwize-music:resume "demo album"

When to use: When user wants to continue working on an existing album.


Instructions

When this skill is invoked with an album name:

Step 1: Find the Album via MCP

  1. Call find_album(name) — fuzzy match by name, slug, or partial (case-insensitive)
  2. If not found: MCP returns available albums — suggest closest match or /bitwize-music:new-album
  3. If multiple matches: list all with paths, ask user which one
  4. If MCP returns stale/missing cache error: call rebuild_state() then retry

Step 2: Get Album Progress

  1. Call get_album_progress(album_slug) — returns track counts by status, completion percentage, and detected workflow phase
  2. Call list_tracks(album_slug) — returns per-track details (status, has_suno_link, sources_verified)

Step 3: Update Session Context

Call update_session(album=album_slug, phase=detected_phase) to set the active album and phase.

Step 4: Determine Current Phase

Based on album and track statuses, identify the workflow phase:

Album StatusTrack StatusesCurrent Phase
ConceptMost "Not Started"Planning - Need to fill in album README and create tracks
Research CompleteSome "Sources Pending"Verification - Need human verification of sources (documentary albums)
Sources VerifiedAll sources verifiedReady to Write - Sources cleared, begin lyrics (documentary albums)
In ProgressMixed, some "Not Started"Writing - Need to complete lyrics (or route instrumental tracks to suno-engineer)
In ProgressSome "Sources Pending"Verification - Need human verification of sources
In ProgressAll have lyricsReady to Generate - Run Ready to Generate checkpoint
In ProgressSome "Generated"Generating - Continue generating on Suno. Check Generation Logs for rejected tracks needing regeneration
In ProgressAll "Generated", none "Final"Review & Approve - Listen to generated tracks, mark keepers with ✓, regenerate rejected ones
CompleteAll "Final"Mastering - Ready to master audio
ReleasedAll "Final"Released - Album is live

Note: Non-documentary albums skip Research Complete and Sources Verified — they go directly from ConceptIn Progress.

Step 5: Report to User

Present a clear status report:

📁 Album: [Album Title]
   Location: {content_root}/artists/{artist}/albums/{genre}/{album}/
   Status: [Album Status]

📊 Progress:
   - Tracks: [X completed / Y total] ([N vocal, M instrumental])
   - Not Started: X
   - In Progress: Y
   - Generated: Z
   - Final: N

📍 Current Phase: [Phase Name]

✅ What's Done:
   - [List completed items]

⏭️ Next Steps:
   1. [Specific action 1]
   2. [Specific action 2]
   3. [Specific action 3]

Ready to continue? Tell me what you'd like to work on.

Step 6: Recommend the Single Best Next Action

Pick ONE clear recommendation from the decision tree below. Don't list 5 options — pick the best one, include the skill name, and be specific about which track.

Decision Tree (evaluate top-to-bottom, first match wins):

Instrumental detection: Check each track's frontmatter for instrumental: true or Track Details table for **Instrumental** | Yes. Instrumental tracks skip the lyrics workflow entirely and go straight to /bitwize-music:suno-engineer.

Album Status = "Concept"
  → "Define the album concept. Run /bitwize-music:album-conceptualizer"

Album Status = "Research Complete"
  → Any tracks Sources Pending?
    YES → "Sources need verification. Run /bitwize-music:verify-sources [album]"
    NO  → Any "Not Started" tracks instrumental?
      YES → "Create Style Box for instrumental track [name]. Use /bitwize-music:suno-engineer"
      NO  → "Ready to write! Pick a track and use /bitwize-music:lyric-writer"

Album has tracks with "Not Started"
  → Is the first not-started track instrumental?
    YES → "Create Style Box for [track]. Use /bitwize-music:suno-engineer directly (instrumental track)"
    NO  → "Write lyrics for [first not-started track]. Use /bitwize-music:lyric-writer"

Album has tracks with "In Progress" (lyrics partially written)
  → "Finish lyrics for [first in-progress track]. Use /bitwize-music:lyric-writer"

Album has tracks with "Sources Pending"
  → "Verify sources for [track]. Run /bitwize-music:verify-sources [album]"

All tracks have lyrics (or Style Box for instrumentals), none generated
  → Mixed album (vocal + instrumental)?
    YES → "All tracks ready! Run /bitwize-music:pronunciation-specialist on vocal tracks, then /bitwize-music:lyric-reviewer, then /bitwize-music:pre-generation-check to validate all gates (instrumental tracks auto-skip lyrics gates)."
    NO  → "All lyrics complete! Style prompts should be ready. Run /bitwize-music:pronunciation-specialist to check for pronunciation risks, then /bitwize-music:lyric-reviewer for final QC, then /bitwize-music:pre-generation-check to validate all gates before generating on Suno."

Some tracks generated, some not
  → Any Generated tracks without ✓ in Generation Log Rating?
    YES → "Track [name] was generated but not approved. Listen and decide:
           - Happy? Mark ✓ in Generation Log and set Status: Final
           - Not happy? Log the reason, then:
             Style issue → /bitwize-music:suno-engineer to revise Style Box
             Lyrics issue → /bitwize-music:lyric-writer to fix, then regenerate
             Bad luck → Regenerate on Suno with same settings (it's non-deterministic)"
    NO  → "Generate [first un-generated track] on Suno. Use /bitwize-music:suno-engineer"

All tracks generated, none Final
  → "All tracks generated! Listen to each track and approve:
     - Mark keepers with ✓ in Generation Log Rating column
     - Reject and regenerate any that don't meet quality standards
     - Once all have ✓, batch-approve:
       Use update_track_field(album_slug, track_slug, 'status', 'Final') for each approved track.
       Once all Final, album advances to Complete automatically."

All tracks generated, some Final
  → Any Generated (non-Final) tracks without ✓?
    YES → "Review track [name] — listen and approve (✓) or regenerate"
    NO  → "All tracks approved! Batch-approve: update_track_field(album_slug, track_slug, 'status', 'Final') for each.
           Then import audio with /bitwize-music:import-audio, then master with /bitwize-music:mastering-engineer"

All tracks Final
  → "All tracks approved! Import audio with /bitwize-music:import-audio, then master with /bitwize-music:mastering-engineer"

Album Status = "Complete"
  → "Album is complete! Release with /bitwize-music:release-director"

Album Status = "Released"
  → "This album is released! Consider /bitwize-music:promo-director for promotional content"
  → Also suggest: "Start a new album? Check /bitwize-music:album-ideas list"

Format the recommendation as:

RECOMMENDED NEXT ACTION:
  [Clear, specific instruction with skill name and track name]

WHY:
  [One sentence explaining why this is the right next step]

When No Album Specified (No Arguments)

If invoked without an album name:

  1. Call get_session() — check last_album from session context, resume that album
  2. If no session context, call list_albums() to find all in-progress albums
  3. Prioritize: closest to completion > unblocked work > last worked on
  4. If no albums exist, suggest /bitwize-music:new-album

Present a multi-album summary if multiple are in progress:

You have X albums. Here's the most actionable:

PRIORITY 1: [album-name] ([genre])
  Status: [status] | Progress: [X/Y tracks]
  → [Recommended action]

Also in progress:
  - [album-2] — [brief status]

Examples

Example 1: Album in Writing Phase

/bitwize-music:resume my-album

📁 Album: My Album
   Location: ~/bitwize-music/artists/bitwize/albums/rock/my-album/
   Status: In Progress

📊 Progress:
   - Tracks: 3 completed / 8 total
   - Not Started: 3
   - In Progress: 2
   - Final: 3

📍 Current Phase: Writing Lyrics

✅ What's Done:
   - Tracks 1-3 have final lyrics
   - Album concept and tracklist defined

⏭️ Next Steps:
   1. Complete lyrics for Track 4 (in progress)
   2. Complete lyrics for Track 5 (in progress)
   3. Write lyrics for Tracks 6-8

Ready to continue? Tell me which track you'd like to work on.

Example 2: Album Ready for Generation

/bitwize-music:resume demo-album

📁 Album: Demo Album
   Location: ~/bitwize-music/artists/bitwize/albums/electronic/demo-album/
   Status: In Progress

📊 Progress:
   - Tracks: 8 / 8 total (all lyrics complete)
   - Final: 8

📍 Current Phase: Ready to Generate

✅ What's Done:
   - All 8 tracks have complete lyrics
   - All lyrics phonetically reviewed
   - Suno Style and Lyrics boxes filled

⏭️ Next Steps:
   1. Run Ready to Generate checkpoint (I'll verify everything)
   2. Start generating on Suno
   3. Log generation attempts

Shall I run the Ready to Generate checkpoint now?

Example 3: Album Not Found

/bitwize-music:resume my-album

❌ Album 'my-album' not found.

Available albums:
- demo-album (electronic) - In Progress
- example-tracks (hip-hop) - Complete

Did you mean one of these? Or use /bitwize-music:new-album to create a new album.

Implementation Notes

  • Use MCP tools - find_album, get_album_progress, list_tracks, update_session instead of reading state.json directly
  • Case-insensitive matching - "Sample-Album" should match "sample-album"
  • Handle missing albums gracefully - List what exists, don't error
  • Be specific about next steps - Don't just say "continue working", say exactly what to do
  • Include full paths - User needs to know where files are located
  • Use emojis sparingly - Only for section headers in the report

Model

Use Sonnet 4.5 - This is a coordination/reporting task, not creative work.

Dépôt GitHub

bitwize-music-studio/claude-ai-music-skills
Chemin: skills/resume
0
ai-musicai-music-toolsaudio-masteringclaudeclaude-codeclaude-code-plugin

Compétences associées

llamaguard

Autre

LlamaGuard est le modèle de Meta, doté de 7 à 8 milliards de paramètres, conçu pour modérer les entrées et sorties des LLM selon six catégories de sécurité comme la violence et les discours haineux. Il offre une précision de 94 à 95 % et peut être déployé avec vLLM, Hugging Face ou Amazon SageMaker. Utilisez cette compétence pour intégrer facilement le filtrage de contenu et des garde-fous de sécurité dans vos applications d'IA.

Voir la compétence

cost-optimization

Autre

Cette compétence de Claude aide les développeurs à optimiser les coûts du cloud grâce au redimensionnement des ressources, aux stratégies d'étiquetage et à l'analyse des dépenses. Elle fournit un cadre pour réduire les dépenses cloud et mettre en œuvre une gouvernance des coûts sur AWS, Azure et GCP. Utilisez-la lorsque vous devez analyser les coûts d'infrastructure, redimensionner les ressources ou respecter des contraintes budgétaires.

Voir la compétence

quantizing-models-bitsandbytes

Autre

Cette compétence quantifie les LLMs en précision 8 bits ou 4 bits à l'aide de bitsandbytes, permettant une réduction de 50 à 75 % de la mémoire utilisée avec une perte de précision minime. Elle est idéale pour exécuter des modèles plus volumineux sur une mémoire GPU limitée ou pour accélérer l'inférence, prenant en charge des formats comme INT8, NF4 et FP4. La compétence s'intègre à HuggingFace Transformers et permet l'entraînement QLoRA ainsi que l'utilisation d'optimiseurs en 8 bits.

Voir la compétence

dispatching-parallel-agents

Autre

Cette compétence Claude déploie plusieurs agents pour enquêter et résoudre simultanément 3 problèmes indépendants ou plus. Elle est conçue pour des scénarios impliquant des défaillances non liées qui peuvent être résolues sans état partagé ni dépendances. La capacité fondamentale est la résolution de problèmes en parallèle, en assignant un agent par domaine problématique indépendant afin de maximiser l'efficacité.

Voir la compétence